UpstageReleased Aug 12, 2026
Proprietary

Solar Pro 4

Its best results are in Academic Reasoning (#100 of 223) and Knowledge Work (#91 of 186); weakest in Programmatic Reasoning (#155 of 187). Abstract Reasoning and Mathematical Reasoning are estimated, because no results are published yet.

Composite IQ ?
102
#127 of 186 ranked models · ahead of 32%
World (Other)'s #12 model, behind mistral-large-4 (118)
Effective cost
$10.40/1M
$0.30 in · $1.20 out
Speed
120tok/s
23s typical response
Context
384Ktokens
384,000 tokens in
Evidence
5of 7 dims
12 of 39 benchmarks

IQ profile

Seven dimensions, averaged equally into Composite IQ
Compare shape with
100 avg human130160● Abstract90#122 · est.● Math93#158 · est.● Academic121#100● Coding98#155● Computer Use89#148● Reliability110#114● Knowledge Work110#91
MeasuredEstimated95% range
#1 in dimensionTop-10 median

IQ dimensions

Ranks are among models scored in each dimension. Ranges are model-based estimates, not calibrated confidence intervals.

DimensionIQPosition · 70–160RankEvidencevs Top 10
90estimated
#122 / 221
0/5
-42.5
93estimated
#158 / 186
0/5
-56.5
121measured
#100 / 223
4/6
-14
98measured
#155 / 187
1/6
-53
89measured
#148 / 186
1/5
-57.5
110measured
#114 / 188
2/6
-25.5
110measured
#91 / 186
4/6
-22
Solar Pro 495% range (model-based, uncalibrated)Top-10 median#1 in dimension100 = average human

Benchmark results

12 of 39 Composite IQ benchmarks have a published result for Solar Pro 4. Benchmarks without a result stay listed so gaps are visible.

Abstract Reasoning IQ 90 · estimatedNo published results0/5
No published results yet for Abstract Reasoning. The 90 shown is estimated from Solar Pro 4's other dimensions and will be replaced as results come in. How estimates work
ARC-AGI-3ARC-AGI-2ARC-AGI-1Chess PuzzlesMystery Game Puzzles
Mathematical Reasoning IQ 93 · estimatedNo published results0/5
No published results yet for Mathematical Reasoning. The 93 shown is estimated from Solar Pro 4's other dimensions and will be replaced as results come in. How estimates work
FrontierMath Tier 4FrontierMath Tier 1-3ProofBench 1.1MathArenaAIME
Academic Reasoning IQ 121 · measuredBest: #56 of 147 on GPQA Diamond4/6
BenchmarkScorePercentileRankLeader · source
GPQA Diamond
Artificial Analysis v5.0.1 · Nov 13
89.1%
#56 / 147
Leader gpt-6-astra · 96.1%
Artificial Analysis v5.0.1 · Nov 13
SciCode
Artificial Analysis v4.3.2 · Oct 27
48.0%
#75 / 167
Leader opus-5.5 · 66.9%
Artificial Analysis v4.3.2 · Oct 27
CritPt
Artificial Analysis v4.3.2 · Oct 27
6.3%
#76 / 159
Leader Muse Wonder Full · 32.9%
Artificial Analysis v4.3.2 · Oct 27
Humanity's Last Exam
Artificial Analysis v4.3.2 · Oct 27
29.8%
#87 / 167
Leader opus-5.5 · 61.4%
Artificial Analysis v4.3.2 · Oct 27
Not yet reported
MMLU-ProMMMU-Pro
Programmatic Reasoning IQ 98 · measuredBest: #109 of 109 on Terminal-Bench 4.01/6
BenchmarkScorePercentileRankLeader · source
Terminal-Bench 4.0
Artificial Analysis v5.0.1 · Nov 13
0.0%
#109 / 109
Leader Kimi K3.1 Pro · 67.2%
Artificial Analysis v5.0.1 · Nov 13
Not yet reported
LiveCodeBenchIOI V2ProgramBenchSWE-rebenchMirrorCode
Computer Use IQ 89 · measuredBest: #77 of 77 on Arena.ai Agent Arena1/5
BenchmarkScorePercentileRankLeader · source
Arena.ai Agent Arena
Arena.ai Agent Arena · Nov 19
-23.9
#77 / 77
Leader fable-5.2 · 27.4
Arena.ai Agent Arena · Nov 19
Not yet reported
BrowseCompOSWorld-VerifiedMCP AtlasAgents' Last Exam
Reliability IQ 110 · measuredBest: #72 of 158 on AA Omniscience2/6
BenchmarkScorePercentileRankLeader · source
AA Omniscience
Artificial Analysis v4.3.2 · Oct 27
0.3
#72 / 158
Leader opus-5.5 · 46.4
Artificial Analysis v4.3.2 · Oct 27
AA Long Context Reasoning v1.1
Artificial Analysis v4.3.2 · Oct 27
70.7%
#107 / 154
Leader kimi-k3 · 88.7%
Artificial Analysis v4.3.2 · Oct 27
Not yet reported
SimpleQA VerifiedBullshitBench v2IFBenchFACTS Grounding
Knowledge Work IQ 110 · measuredBest: #53 of 116 on AA-Briefcase v1.1 Elo4/6
BenchmarkScorePercentileRankLeader · source
AA-Briefcase v1.1 Elo
Artificial Analysis v5.0.1 · Nov 13
1,093
#53 / 116
Leader sonnet-5.5 · 1,816
Artificial Analysis v5.0.1 · Nov 13
GDPval-AA v2.1 normalized score
Artificial Analysis v4.3.2 · Oct 23
30.3%
#61 / 124
Leader opus-5.5 · 68.3%
Artificial Analysis v4.3.2 · Oct 23
AutomationBench-AA
Artificial Analysis v5.0.1 · Nov 13
13.6%
#70 / 112
Leader Gemini 4 Carbon · 77.8%
Artificial Analysis v5.0.1 · Nov 13
GDP.pdf
Artificial Analysis v5.0.1 · Nov 13
7.0%
#75 / 111
Leader Muse Wonder Full · 32.6%
Artificial Analysis v5.0.1 · Nov 13
Not yet reported
AA-AnalystAgentEnterpriseOps-Gym-AA

Release & availability

Facts from the publisher's announcement and documentation.

Released
Developer
Upstage
Access
Available through the Upstage API and third-party API providers
Weights
ProprietaryHosted access only
Inputs
text
Context
384,000 tokens
API price
$0.30 in · $1.20 outper 1M tokens · full pricing

What Upstage says

  • Solar Pro 4 is Upstage's proprietary flagship successor to Solar Pro 3.
  • It supports a 384k-token context window and up to 256k output tokens.

Limitations

  • The launch is text-only and model weights are not available.

Publisher statements, not AI IQ measurements.

About this score

How much of Solar Pro 4's IQ is measured, and how the rest is estimated.

12/ 39benchmark results
5/ 7dimensions measured
2estimated
0/5Abstract
0/5Math
4/6Academic
1/6Coding
1/5Computer Use
2/6Reliability
4/6Knowledge Work
Bar height = share of benchmarks with resultsEstimatedNo score
Measured results are shown as published

Every score in the benchmark tables comes from its source unchanged. Composite IQ averages all seven dimensions equally.

Estimates fill the gaps

Abstract Reasoning and Mathematical Reasoning have no published results yet, so their IQ is estimated from Solar Pro 4's other dimensions. Missing benchmarks can also draw on its predecessor, Solar Pro 3; a predecessor is an estimation input, not evidence that this model has the predecessor's measured results.

A missing result is not a zero

Benchmark versions and evaluation harnesses can differ between sources, so small gaps between models may not be meaningful.

Read the scoring methodology →

Compare Solar Pro 4

Side-by-side pages for the models it is most often weighed against.

Solar Pro 4Comparison model
Browse all comparisons →

Frequently asked questions

Quick answers about Solar Pro 4, drawn from the same data as the rest of this page.

How AI IQ scores models →Get this data from the API →
What is the IQ of Solar Pro 4?
Solar Pro 4 has an estimated IQ of 102, which ranks #127 out of 186 public models on the AI IQ leaderboard. The score combines seven equally weighted reasoning dimensions using measured benchmark results and conservative estimates for missing coverage.
What is Solar Pro 4 best at?
Among the dimensions with published benchmark results, Solar Pro 4 ranks highest in Academic Reasoning (IQ 121, #100 of 223) and Knowledge Work (IQ 110, #91 of 186). Its lowest-ranked measured dimension is Programmatic Reasoning (IQ 98, #155 of 187).
What is the EQ of Solar Pro 4?
Solar Pro 4 doesn't have an EQ score yet. EQ (emotional reasoning) is calculated from EQ-Bench 4 and AttuneBench, and neither has published a result for Solar Pro 4. AI IQ doesn't estimate EQ from other benchmarks, so the score appears once one of those results is published.
Which benchmarks has Solar Pro 4 been tested on?
Solar Pro 4 has published results on 12 of the 39 benchmarks behind AI IQ's Composite IQ, including GPQA Diamond, SciCode, AA Omniscience, AA-Briefcase v1.1 Elo and CritPt.
How much does Solar Pro 4 cost?
Solar Pro 4 costs $0.30 per million input tokens and $1.20 per million output tokens. Adjusted for how many tokens it actually consumes on real tasks, its effective cost works out to $10.40 per million input-plus-output tokens.
Is Solar Pro 4 open source?
No — Solar Pro 4 is a proprietary model. Its weights are not publicly available, so it is used through hosted APIs and products.
Who developed Solar Pro 4?
Solar Pro 4 was developed by Upstage, based in South Korea.
When was Solar Pro 4 released?
Solar Pro 4 was publicly released on August 12, 2026.
How large is the context window of Solar Pro 4?
Solar Pro 4 supports a context window of 384,000 tokens — the amount of text it can consider in a single request.
How fast is Solar Pro 4?
In independent measurements, Solar Pro 4 generates about 120 tokens per second, with a typical response time of about 23 seconds.